| 无约束最优化中两种改进共轭梯度法的收敛性证明 |
| |
| 引用本文： | 周安娃,范浩,黄青群. 无约束最优化中两种改进共轭梯度法的收敛性证明[J]. 桂林电子科技大学学报, 2011, 31(1): 37-40 |
| |
| 作者姓名： | 周安娃 范浩 黄青群 |
| |
| 作者单位： | 桂林电子科技大学,数学与计算科学学院,广西,桂林,541004 |
| |
| 基金项目： | 国家自然科学基金，广西高校优秀人才资助项目 |
| |
| 摘 要： | 对于无约束优化中已提出的两种改进共轭梯度算法:改进的DY算法(MDY)和新的混合HS-DY算法(NH),证明了其在Wolfe线搜索下的全局收敛性.证明中的关键技巧是利用DY算法公式的一个等价公式,也正是由于该策略的运用,使得证明更为简化,进而得到了上述两种改进的共轭梯度法的全局收敛性.
|
| 关 键 词： | 共轭梯度法 下降方向 全局收敛性 |
The global convergence of two modified conjugate gradiet methods for optimization |
| |
| Zhou Anwa,Fan Hao,Huang Qingqun. The global convergence of two modified conjugate gradiet methods for optimization[J]. Journal of Guilin University of Electronic Technology, 2011, 31(1): 37-40 |
| |
| Authors: | Zhou Anwa Fan Hao Huang Qingqun |
| |
| Affiliation: | Zhou Anwa,Fan Hao,Huang Qingqun(School of Mathematics and Computational Science,Guilin University of Electronic Technology,Guilin 541004,China) |
| |
| Abstract: | Considering two modified conjugare gradient methods which had been presented: the modified DY method and the new hybrid HS-DY method for unconvex function,we explore the global convergence of these two methods with the Wolfe line search.It is very important to use a equivalence formula of DY method for the proof,which makes it simplify and shows that these modified methods are both globally convergent. |
| |
| Keywords: | conjugate gradient method descent direction global convergence |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
|
